Application of digital image correlation to full-field measurement of shrinkage strain of dental composites

Objectives: Polymerization shrinkage of dental composites remains a major concern in restorative dentistry because it can lead to micro-cracking of the tooth and debonding at the tooth-restoration interface. The aim of this study was to measure the full-field polymerization shrinkage of dental composites using the optical digital image correlation (DIC) method and to evaluate how the measurement is influenced by the factors in experiment setup and image analysis. Methods: Four commercial dental composites, Premise Dentine, Z100, Z250 and Tetric EvoCeram, were tested. Composite was first placed into a slot mould to form a bar specimen with rectangular-section of 4 mm×2 mm, followed by the surface painting to create irregular speckles. Curing was then applied at one end of the specimen while the other part were covered against curing light for simulating the clinical curing condition of composite in dental cavity. The painted surface was recorded by a charge-coupled device (CCD) camera before and after curing. Subsequently, the volumetric shrinkage of the specimen was calculated with specialist DIC software based on image cross correlation. In addition, a few factors that may influence the measuring accuracy, including the subset window size, speckle size, illumination light and specimen length, were also evaluated. Results: The volumetric shrinkage of the specimen generally decreases with increasing distance from the irradiated surface with a conspicuous exception being the composite Premise Dentine as its maximum shrinkage occurred at a subsurface distance of about 1 mm instead of the irradiated surface. Z100 had the greatest maximum shrinkage strain, followed by Z250, Tetric EvoCeram and then Premise Dentine. Larger subset window size made the shrinkage strain contour smoother. But the cost was that some details in the heterogeneity of the material were lost. Very small subset window size resulted in a lot of noise in the data, making it difficult to discern the general pattern in the strain distribution. Speckle size did not seem to have obvious effect on the volumetric shrinkage strain along specimen length; however, larger speckles resulted in higher level of noise or heterogeneity in the shrinkage distribution. Compared with bright illumination, dimmer lighting produced larger standard deviations in the measured shrinkage, indicating a higher level of noise. The longer the specimen, the greater was the rate of reduction with distance from the irradiated surface, especially for the longitudinal strain. Significance: The image correlation method is capable of producing full-field polymerization shrinkage of dental composites. The accuracy of the measurements relies on selection of optimal parameters in experimental setup and DIC analysis.